VideoChat: Chat-Centric Video Understanding

Benchmark Model Rank Results
question-answering-on-next-qa-open-endedVideoChat#2Accuracy: 56.6Confidence Score: 3.2
video-based-generative-performanceVideo Chat#21mean: 2.29Correctness of Information: 2.23Detail Orientation: 2.50
video-based-generative-performance-1Video Chat#15gpt-score: 2.32
video-based-generative-performance-2Video Chat#15gpt-score: 2.24
video-based-generative-performance-3Video Chat#16gpt-score: 2.53
video-based-generative-performance-4Video Chat#15gpt-score: 2.50
video-based-generative-performance-5Video Chat#17gpt-score: 1.94
video-question-answering-on-activitynet-qaVideo Chat#29Accuracy: 26.5Confidence score: 2.2
video-question-answering-on-mvbenchVideoChat#18Avg.: 35.5
zero-shot-video-question-answer-on-vnbenchVideoChat2#7Accuracy: 12.4
zeroshot-video-question-answer-on-activitynetVideo Chat#26Accuracy: 26.5Confidence Score: 2.2
zeroshot-video-question-answer-on-msrvtt-qaVideo Chat-7B#27Accuracy: 45.0Confidence Score: 2.5
zeroshot-video-question-answer-on-msvd-qaVideo Chat-7B#25Accuracy: 56.3Confidence Score: 2.8
zeroshot-video-question-answer-on-tgif-qaVideo Chat-7B#14Accuracy: 34.4Confidence Score: 2.3